Water Quality Prediction Project

A Data Science project by Aditya Kumar, Aditya Nayak & Kunal Inglunkar

import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import seaborn as sns
import altair as alt # Interactive data viz package cmd: pip install altair
import pygwalker as pyg #Access tableau feature in python cmd: pip install pygwalker
from skimpy import skim, generate_test_data #Gives overall stats cmd: pip install skimpy
from summarytools import dfSummary #combination of info & describe fuctions. cmd: pip install summarytools
import sidetable # Better value_counts function cmd: pip install sidetable
import warnings
warnings.filterwarnings("ignore")

About Dataset

Context:

Access to safe drinking-water is essential to health, a basic human right and a component of effective policy for health protection. This is important as a health and development issue at a national, regional and local level. In some regions, it has been shown that investments in water supply and sanitation can yield a net economic benefit, since the reductions in adverse health effects and health care costs outweigh the costs of undertaking the interventions.

Content

The water_potability.csv file contains water quality metrics for 3276 different water bodies.

1. pH value:

PH is an important parameter in evaluating the acid–base balance of water. It is also the indicator of acidic or alkaline condition of water status. WHO has recommended maximum permissible limit of pH from 6.5 to 8.5. The current investigation ranges were 6.52–6.83 which are in the range of WHO standards.



2. Hardness:

Hardness is mainly caused by calcium and magnesium salts. These salts are dissolved from geologic deposits through which water travels. The length of time water is in contact with hardness producing material helps determine how much hardness there is in raw water. Hardness was originally defined as the capacity of water to precipitate soap caused by Calcium and Magnesium.



3. Solids (Total dissolved solids - TDS):

Water has the ability to dissolve a wide range of inorganic and some organic minerals or salts such as potassium, calcium, sodium, bicarbonates, chlorides, magnesium, sulfates etc. These minerals produced un-wanted taste and diluted color in appearance of water. This is the important parameter for the use of water. The water with high TDS value indicates that water is highly mineralized. Desirable limit for TDS is 500 mg/l and maximum limit is 1000 mg/l which prescribed for drinking purpose.



4. Chloramines:

Chlorine and chloramine are the major disinfectants used in public water systems. Chloramines are most commonly formed when ammonia is added to chlorine to treat drinking water. Chlorine levels up to 4 milligrams per liter (mg/L or 4 parts per million (ppm)) are considered safe in drinking water.



5. Sulfate:

Sulfates are naturally occurring substances that are found in minerals, soil, and rocks. They are present in ambient air, groundwater, plants, and food. The principal commercial use of sulfate is in the chemical industry. Sulfate concentration in seawater is about 2,700 milligrams per liter (mg/L). It ranges from 3 to 30 mg/L in most freshwater supplies, although much higher concentrations (1000 mg/L) are found in some geographic locations.



6. Conductivity:

Pure water is not a good conductor of electric current rather’s a good insulator. Increase in ions concentration enhances the electrical conductivity of water. Generally, the amount of dissolved solids in water determines the electrical conductivity. Electrical conductivity (EC) actually measures the ionic process of a solution that enables it to transmit current. According to WHO standards, EC value should not exceeded 400 μS/cm.



7. Organic_carbon:

Total Organic Carbon (TOC) in source waters comes from decaying natural organic matter (NOM) as well as synthetic sources. TOC is a measure of the total amount of carbon in organic compounds in pure water. According to US EPA < 2 mg/L as TOC in treated / drinking water, and < 4 mg/Lit in source water which is use for treatment.



8. Trihalomethanes:

THMs are chemicals which may be found in water treated with chlorine. The concentration of THMs in drinking water varies according to the level of organic material in the water, the amount of chlorine required to treat the water, and the temperature of the water that is being treated. THM levels up to 80 ppm is considered safe in drinking water.



9. Turbidity:

The turbidity of water depends on the quantity of solid matter present in the suspended state. It is a measure of light emitting properties of water and the test is used to indicate the quality of waste discharge with respect to colloidal matter. The mean turbidity value obtained for Wondo Genet Campus (0.98 NTU) is lower than the WHO recommended value of 5.00 NTU.



10. Potability:

Indicates if water is safe for human consumption where 1 means Potable and 0 means Not potable.

1. Exploratory Data Analysis

1.1 Data Cleaning

df = pd.read_csv("water_potability.csv")
df.head()

	ph	Hardness	Solids	Chloramines	Sulfate	Conductivity	Organic_carbon	Trihalomethanes	Turbidity	Potability
0	NaN	204.890455	20791.318981	7.300212	368.516441	564.308654	10.379783	86.990970	2.963135	0
1	3.716080	129.422921	18630.057858	6.635246	NaN	592.885359	15.180013	56.329076	4.500656	0
2	8.099124	224.236259	19909.541732	9.275884	NaN	418.606213	16.868637	66.420093	3.055934	0
3	8.316766	214.373394	22018.417441	8.059332	356.886136	363.266516	18.436524	100.341674	4.628771	0
4	9.092223	181.101509	17978.986339	6.546600	310.135738	398.410813	11.558279	31.997993	4.075075	0

df.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 3276 entries, 0 to 3275
Data columns (total 10 columns):
 #   Column           Non-Null Count  Dtype  
---  ------           --------------  -----  
 0   ph               2785 non-null   float64
 1   Hardness         3276 non-null   float64
 2   Solids           3276 non-null   float64
 3   Chloramines      3276 non-null   float64
 4   Sulfate          2495 non-null   float64
 5   Conductivity     3276 non-null   float64
 6   Organic_carbon   3276 non-null   float64
 7   Trihalomethanes  3114 non-null   float64
 8   Turbidity        3276 non-null   float64
 9   Potability       3276 non-null   int64  
dtypes: float64(9), int64(1)
memory usage: 256.1 KB

df.describe()

	ph	Hardness	Solids	Chloramines	Sulfate	Conductivity	Organic_carbon	Trihalomethanes	Turbidity	Potability
count	2785.000000	3276.000000	3276.000000	3276.000000	2495.000000	3276.000000	3276.000000	3114.000000	3276.000000	3276.000000
mean	7.080795	196.369496	22014.092526	7.122277	333.775777	426.205111	14.284970	66.396293	3.966786	0.390110
std	1.594320	32.879761	8768.570828	1.583085	41.416840	80.824064	3.308162	16.175008	0.780382	0.487849
min	0.000000	47.432000	320.942611	0.352000	129.000000	181.483754	2.200000	0.738000	1.450000	0.000000
25%	6.093092	176.850538	15666.690297	6.127421	307.699498	365.734414	12.065801	55.844536	3.439711	0.000000
50%	7.036752	196.967627	20927.833607	7.130299	333.073546	421.884968	14.218338	66.622485	3.955028	0.000000
75%	8.062066	216.667456	27332.762127	8.114887	359.950170	481.792304	16.557652	77.337473	4.500320	1.000000
max	14.000000	323.124000	61227.196008	13.127000	481.030642	753.342620	28.300000	124.000000	6.739000	1.000000

skim(df)

╭──────────────────────────────────────────────── skimpy summary ─────────────────────────────────────────────────╮
│          Data Summary                Data Types                                                                 │
│ ┏━━━━━━━━━━━━━━━━━━━┳━━━━━━━━┓ ┏━━━━━━━━━━━━━┳━━━━━━━┓                                                          │
│ ┃ dataframe         ┃ Values ┃ ┃ Column Type ┃ Count ┃                                                          │
│ ┡━━━━━━━━━━━━━━━━━━━╇━━━━━━━━┩ ┡━━━━━━━━━━━━━╇━━━━━━━┩                                                          │
│ │ Number of rows    │ 3276   │ │ float64     │ 9     │                                                          │
│ │ Number of columns │ 10     │ │ int32       │ 1     │                                                          │
│ └───────────────────┴────────┘ └─────────────┴───────┘                                                          │
│                                                     number                                                      │
│ ┏━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━┳━━━━━━━━┳━━━━━━━━━┳━━━━━━━━┳━━━━━━━━┳━━━━━━━━━┳━━━━━━━━━┳━━━━━━━━━┳━━━━━━━━━┓  │
│ ┃ column_name           ┃ NA    ┃ NA %   ┃ mean    ┃ sd     ┃ p0     ┃ p25     ┃ p75     ┃ p100    ┃ hist    ┃  │
│ ┡━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━╇━━━━━━━━╇━━━━━━━━━╇━━━━━━━━╇━━━━━━━━╇━━━━━━━━━╇━━━━━━━━━╇━━━━━━━━━╇━━━━━━━━━┩  │
│ │ ph                    │   490 │     15 │     7.1 │    1.6 │      0 │     6.1 │     8.1 │      14 │   ▁██▁  │  │
│ │ Hardness              │     0 │      0 │     200 │     33 │     47 │     180 │     220 │     320 │   ▁▅█▂  │  │
│ │ Solids                │     0 │      0 │   22000 │   8800 │    320 │   16000 │   27000 │   61000 │  ▁█▇▂   │  │
│ │ Chloramines           │     0 │      0 │     7.1 │    1.6 │   0.35 │     6.1 │     8.1 │      13 │   ▁▆█▂  │  │
│ │ Sulfate               │   780 │     24 │     330 │     41 │    130 │     310 │     360 │     480 │    ▃█▃  │  │
│ │ Conductivity          │     0 │      0 │     430 │     81 │    180 │     370 │     480 │     750 │   ▅█▅▁  │  │
│ │ Organic_carbon        │     0 │      0 │      14 │    3.3 │    2.2 │      12 │      17 │      28 │   ▂█▆▁  │  │
│ │ Trihalomethanes       │   160 │    4.9 │      66 │     16 │   0.74 │      56 │      77 │     120 │   ▁▆█▂  │  │
│ │ Turbidity             │     0 │      0 │       4 │   0.78 │    1.4 │     3.4 │     4.5 │     6.7 │   ▃█▇▂  │  │
│ │ Potability            │     0 │      0 │    0.39 │   0.49 │      0 │       0 │       1 │       1 │ █    ▅  │  │
│ └───────────────────────┴───────┴────────┴─────────┴────────┴────────┴─────────┴─────────┴─────────┴─────────┘  │
╰────────────────────────────────────────────────────── End ──────────────────────────────────────────────────────╯

From the following description of the dataset we observe:

• That there are missing values in ph, Sulfate & Trihalomethanes.
  
• Potability is a categorical variable which tells if the water is potable(1) or not potable(0).
  
• Thus, converting the potability column from int to charecter.
  

def pot_label(p):
    if(p == 0):
        return 'No'
    else:
        return 'Yes'
    
df.Potability = df.Potability.apply(pot_label)
df.head()

	ph	Hardness	Solids	Chloramines	Sulfate	Conductivity	Organic_carbon	Trihalomethanes	Turbidity	Potability
0	NaN	204.890455	20791.318981	7.300212	368.516441	564.308654	10.379783	86.990970	2.963135	No
1	3.716080	129.422921	18630.057858	6.635246	NaN	592.885359	15.180013	56.329076	4.500656	No
2	8.099124	224.236259	19909.541732	9.275884	NaN	418.606213	16.868637	66.420093	3.055934	No
3	8.316766	214.373394	22018.417441	8.059332	356.886136	363.266516	18.436524	100.341674	4.628771	No
4	9.092223	181.101509	17978.986339	6.546600	310.135738	398.410813	11.558279	31.997993	4.075075	No

dfSummary(df)

Data Frame Summary
df
Dimensions: 3,276 x 10
Duplicates: 0

No	Variable	Stats / Values	Freqs / (% of Valid)	Graph	Missing
1	ph [float64]	Mean (sd) : 7.1 (1.6) min < med < max: 0.0 < 7.0 < 14.0 IQR (CV) : 2.0 (4.4)	2,785 distinct values		491 (15.0%)
2	Hardness [float64]	Mean (sd) : 196.4 (32.9) min < med < max: 47.4 < 197.0 < 323.1 IQR (CV) : 39.8 (6.0)	3,276 distinct values		0 (0.0%)
3	Solids [float64]	Mean (sd) : 22014.1 (8768.6) min < med < max: 320.9 < 20927.8 < 61227.2 IQR (CV) : 11666.1 (2.5)	3,276 distinct values		0 (0.0%)
4	Chloramines [float64]	Mean (sd) : 7.1 (1.6) min < med < max: 0.4 < 7.1 < 13.1 IQR (CV) : 2.0 (4.5)	3,276 distinct values		0 (0.0%)
5	Sulfate [float64]	Mean (sd) : 333.8 (41.4) min < med < max: 129.0 < 333.1 < 481.0 IQR (CV) : 52.3 (8.1)	2,495 distinct values		781 (23.8%)
6	Conductivity [float64]	Mean (sd) : 426.2 (80.8) min < med < max: 181.5 < 421.9 < 753.3 IQR (CV) : 116.1 (5.3)	3,276 distinct values		0 (0.0%)
7	Organic_carbon [float64]	Mean (sd) : 14.3 (3.3) min < med < max: 2.2 < 14.2 < 28.3 IQR (CV) : 4.5 (4.3)	3,276 distinct values		0 (0.0%)
8	Trihalomethanes [float64]	Mean (sd) : 66.4 (16.2) min < med < max: 0.7 < 66.6 < 124.0 IQR (CV) : 21.5 (4.1)	3,114 distinct values		162 (4.9%)
9	Turbidity [float64]	Mean (sd) : 4.0 (0.8) min < med < max: 1.4 < 4.0 < 6.7 IQR (CV) : 1.1 (5.1)	3,276 distinct values		0 (0.0%)
10	Potability [object]	1. No 2. Yes	1,998 (61.0%) 1,278 (39.0%)		0 (0.0%)

#Drop Viz?
sns.set_style("whitegrid")
a = sns.catplot('Potability',kind = 'count',data = df ,palette='autumn')
a.set(xlabel ="Water Potability?", ylabel = "count", title ='Potability of water')

<seaborn.axisgrid.FacetGrid at 0x20d4ba3f100>

In order to fill the missing values of pH, Sulfate & Trihalomethane columns, we will have to have a look at the distribution & based on that the appropiate approximation can be made.

#Drop viz?
chart = alt.Chart(df).mark_bar().encode(
    alt.X('ph:Q',bin = True,axis = alt.Axis(title="pH value")),
    alt.Y('count()', axis = alt.Axis(title="count"))).properties(title="Distribution of pH values")
chart

#Drop viz?
chart1 = alt.Chart(df).mark_bar().encode(
    alt.X('Sulfate:Q',bin = True,axis = alt.Axis(title="Sulfate content (mg\L)")),
    alt.Y('count()', axis = alt.Axis(title="count"))).properties(title="Distribution of Sulfate values")
chart1

#Drop viz?
chart2 = alt.Chart(df).mark_bar().encode(
    alt.X('Trihalomethanes:Q',bin = True,axis = alt.Axis(title="Trihalomethanes content (μg/L)")),
    alt.Y('count()', axis = alt.Axis(title="count"))).properties(title="Distribution of Trihalomethane values")
chart2

From the histograms of pH & Trihalomethanes, we can infer that the variables are normally distributed and Sulfates are somewhat left-skewed. Hence can take the mean values of pH & Trihalomethanes and the meadian of Sulfates variables.

df.ph.fillna(df.ph.mean(),inplace=True)
df.Sulfate.fillna(df.Sulfate.median(),inplace=True)
df.Trihalomethanes.fillna(df.Trihalomethanes.mean(),inplace=True)
df.to_csv('Cleaned_data.csv')
df.head()

	ph	Hardness	Solids	Chloramines	Sulfate	Conductivity	Organic_carbon	Trihalomethanes	Turbidity	Potability
0	7.080795	204.890455	20791.318981	7.300212	368.516441	564.308654	10.379783	86.990970	2.963135	No
1	3.716080	129.422921	18630.057858	6.635246	333.073546	592.885359	15.180013	56.329076	4.500656	No
2	8.099124	224.236259	19909.541732	9.275884	333.073546	418.606213	16.868637	66.420093	3.055934	No
3	8.316766	214.373394	22018.417441	8.059332	356.886136	363.266516	18.436524	100.341674	4.628771	No
4	9.092223	181.101509	17978.986339	6.546600	310.135738	398.410813	11.558279	31.997993	4.075075	No

skim(df)

╭──────────────────────────────────────────────── skimpy summary ─────────────────────────────────────────────────╮
│          Data Summary                Data Types                                                                 │
│ ┏━━━━━━━━━━━━━━━━━━━┳━━━━━━━━┓ ┏━━━━━━━━━━━━━┳━━━━━━━┓                                                          │
│ ┃ dataframe         ┃ Values ┃ ┃ Column Type ┃ Count ┃                                                          │
│ ┡━━━━━━━━━━━━━━━━━━━╇━━━━━━━━┩ ┡━━━━━━━━━━━━━╇━━━━━━━┩                                                          │
│ │ Number of rows    │ 3276   │ │ float64     │ 9     │                                                          │
│ │ Number of columns │ 10     │ │ string      │ 1     │                                                          │
│ └───────────────────┴────────┘ └─────────────┴───────┘                                                          │
│                                                     number                                                      │
│ ┏━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━┳━━━━━━━━┳━━━━━━━━━━┳━━━━━━━━┳━━━━━━━━┳━━━━━━━━━┳━━━━━━━━━┳━━━━━━━━━┳━━━━━━━━━┓  │
│ ┃ column_name           ┃ NA   ┃ NA %   ┃ mean     ┃ sd     ┃ p0     ┃ p25     ┃ p75     ┃ p100    ┃ hist    ┃  │
│ ┡━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━╇━━━━━━━━╇━━━━━━━━━━╇━━━━━━━━╇━━━━━━━━╇━━━━━━━━━╇━━━━━━━━━╇━━━━━━━━━╇━━━━━━━━━┩  │
│ │ ph                    │    0 │      0 │      7.1 │    1.5 │      0 │     6.3 │     7.9 │      14 │   ▁▆█▁  │  │
│ │ Hardness              │    0 │      0 │      200 │     33 │     47 │     180 │     220 │     320 │   ▁▅█▂  │  │
│ │ Solids                │    0 │      0 │    22000 │   8800 │    320 │   16000 │   27000 │   61000 │  ▁█▇▂   │  │
│ │ Chloramines           │    0 │      0 │      7.1 │    1.6 │   0.35 │     6.1 │     8.1 │      13 │   ▁▆█▂  │  │
│ │ Sulfate               │    0 │      0 │      330 │     36 │    130 │     320 │     350 │     480 │    ▂█▂  │  │
│ │ Conductivity          │    0 │      0 │      430 │     81 │    180 │     370 │     480 │     750 │   ▅█▅▁  │  │
│ │ Organic_carbon        │    0 │      0 │       14 │    3.3 │    2.2 │      12 │      17 │      28 │   ▂█▆▁  │  │
│ │ Trihalomethanes       │    0 │      0 │       66 │     16 │   0.74 │      57 │      77 │     120 │   ▁▅█▂  │  │
│ │ Turbidity             │    0 │      0 │        4 │   0.78 │    1.4 │     3.4 │     4.5 │     6.7 │   ▃█▇▂  │  │
│ └───────────────────────┴──────┴────────┴──────────┴────────┴────────┴─────────┴─────────┴─────────┴─────────┘  │
│                                                     string                                                      │
│ ┏━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━┳━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━┓  │
│ ┃ column_name               ┃ NA      ┃ NA %       ┃ words per row                ┃ total words              ┃  │
│ ┡━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━╇━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━┩  │
│ │ Potability                │       0 │          0 │                            1 │                     3300 │  │
│ └───────────────────────────┴─────────┴────────────┴──────────────────────────────┴──────────────────────────┘  │
╰────────────────────────────────────────────────────── End ──────────────────────────────────────────────────────╯

We now have a complete dataset & can begin our analysis.

1.2 Data Visualization

sns.boxplot(data=df, x="ph",y = 'Potability')
plt.title("Distribution of pH based on Potability of water")
plt.xlabel("pH")

Text(0.5, 0, 'pH')

fig, (ax_box, ax_violin) = plt.subplots(nrows=1, ncols=2, figsize=(13, 5))
sns.boxplot(x="ph", y="Potability", data=df, ax=ax_box)
sns.violinplot(x="ph", y="Potability", data=df, ax=ax_violin)
ax_box.set_title("Boxplot of pH distribution")
ax_violin.set_title("Violinplot of pH distribution")

Text(0.5, 1.0, 'Violinplot of pH distribution')

fig, (ax_box1, ax_violin1) = plt.subplots(nrows=1, ncols=2, figsize=(15, 5.5))
sns.boxplot(data=df[['Chloramines','Organic_carbon','Turbidity']],ax=ax_box1,orient='h')
sns.violinplot(data=df[['Chloramines','Organic_carbon','Turbidity']],ax=ax_violin1,orient='h')
ax_box1.set_title("Boxplot of Chloramines(ppm), Organic Carbon(ppm) & Turbidity(NTU)")
ax_violin1.set_title("Violinplot of Chloramines(ppm), Organic Carbon(ppm) & Turbidity(NTU)")

Text(0.5, 1.0, 'Violinplot of Chloramines(ppm), Organic Carbon(ppm) & Turbidity(NTU)')

fig, (ax_box2, ax_violin2) = plt.subplots(nrows=1, ncols=2, figsize=(15, 5.5))
sns.boxplot(data=df[['Sulfate','Conductivity']],orient="h",ax=ax_box2)
sns.violinplot(data=df[['Sulfate','Conductivity']],ax=ax_violin2,orient='h')
ax_box2.set_title("Boxplot of Sulfates(mg/L) & Conductivity(μS/cm)")
ax_violin2.set_title("Violinplot of Sulfates(mg/L) & Conductivity(μS/cm)")

Text(0.5, 1.0, 'Violinplot of Sulfates(mg/L) & Conductivity(μS/cm)')

fig, (ax_box3, ax_violin3) = plt.subplots(nrows=1, ncols=2, figsize=(15, 5.5))
sns.boxplot(data=df.Solids,orient="h",ax=ax_box3)
sns.violinplot(data = df.Solids,orient="h",ax=ax_violin3)
ax_box3.set_title("Boxplot distribution of Total Dissolved Solids(ppm)")
ax_violin3.set_title("Violinplot distribution of Total Dissolved Solids(ppm)")
ax_box3.set_xlabel("Sulfate Amount(ppm)")
ax_violin3.set_xlabel("Sulfate Amount(ppm)")

Text(0.5, 0, 'Sulfate Amount(ppm)')

alt.Chart(df).mark_circle(size=60).encode(
    x = alt.X('ph',axis=alt.Axis(title="pH")),
    y = alt.Y('Hardness',axis=alt.Axis(title="Hardness(mg/L)")),
    color=alt.Color('Potability:O',scale = alt.Scale(scheme='magma')),
    tooltip=['Solids:Q','Chloramines:Q','Sulfate:Q','Conductivity:Q','Organic_carbon:Q','Trihalomethanes:Q','Turbidity:Q']
).properties(title="pH vs. Hardness").interactive()
#Double tap graph to re-center

alt.Chart(df).mark_circle(size=60).encode(
    x = alt.X('ph',axis=alt.Axis(title = "pH")),
    y = alt.Y('Solids',axis = alt.Axis(title = 'TDS amount(ppm)')),
    color = alt.Color('Potability:O',scale = alt.Scale(scheme='set1')),
    tooltip = ['Hardness:Q','Chloramines:Q','Sulfate:Q','Conductivity:Q','Organic_carbon:Q','Trihalomethanes:Q','Turbidity:Q']
).properties(title="pH vs. Total Dissolved Solids").interactive()

alt.Chart(df).mark_circle(size=60).encode(
    x = alt.X('ph',axis=alt.Axis(title = "pH")),
    y = alt.Y('Chloramines',axis = alt.Axis(title = 'Chloramine amount(ppm)')),
    color = alt.Color('Potability:O',scale = alt.Scale(scheme='category10')),
    tooltip = ['Hardness:Q','Solids:Q','Sulfate:Q','Conductivity:Q','Organic_carbon:Q','Trihalomethanes:Q','Turbidity:Q']
).properties(title="pH vs. Chloramines").interactive()

alt.Chart(df).mark_circle(size=60).encode(
    x = alt.X('ph',axis=alt.Axis(title = "pH")),
    y = alt.Y('Sulfate',axis = alt.Axis(title = 'Sulfate amount(mg/L)')),
    color = alt.Color('Potability:O',scale = alt.Scale(scheme='dark2')),
    tooltip = ['Hardness:Q','Solids:Q','Chloramines:Q','Conductivity:Q','Organic_carbon:Q','Trihalomethanes:Q','Turbidity:Q']
).properties(title="pH vs. Sulfates").interactive()

alt.Chart(df).mark_circle(size=60).encode(
    x = alt.X('ph',axis=alt.Axis(title = "pH")),
    y = alt.Y('Conductivity',axis = alt.Axis(title = 'Conductivity of water(μS/cm)')),
    color = alt.Color('Potability:O',scale = alt.Scale(scheme='viridis')),
    tooltip = ['Hardness:Q','Solids:Q','Chloramines:Q','Sulfates:Q','Organic_carbon:Q','Trihalomethanes:Q','Turbidity:Q']
).properties(title="pH vs. Conductivity").interactive()

alt.Chart(df).mark_circle(size=60).encode(
    x = alt.X('ph',axis=alt.Axis(title = "pH")),
    y = alt.Y('Organic_carbon',axis = alt.Axis(title = 'Amount of Organic Carbon(ppm)')),
    color = alt.Color('Potability:O',scale = alt.Scale(scheme='set2')),
    tooltip = ['Hardness:Q','Solids:Q','Chloramines:Q','Sulfates:Q','Conductivity:Q','Trihalomethanes:Q','Turbidity:Q']
).properties(title="pH vs. Organic Carbon").interactive()

alt.Chart(df).mark_circle(size=60).encode(
    x = alt.X('ph',axis=alt.Axis(title = "pH")),
    y = alt.Y('Trihalomethanes',axis = alt.Axis(title = 'Trihalomethanes amount(μg/L)')),
    color = alt.Color('Potability:O',scale = alt.Scale(scheme='magma')),
    tooltip = ['Hardness:Q','Solids:Q','Chloramines:Q','Sulfates:Q','Conductivity:Q','Organic_carbon:Q','Turbidity:Q']
).properties(title="pH vs. Trihalomethanes").interactive()

alt.Chart(df).mark_circle(size=60).encode(
    x = alt.X('ph',axis=alt.Axis(title = "pH")),
    y = alt.Y('Turbidity',axis = alt.Axis(title = 'Turbidity(NTU)')),
    color = alt.Color('Potability:O',scale = alt.Scale(scheme='darkmulti')),
    tooltip = ['Hardness:Q','Solids:Q','Chloramines:Q','Sulfates:Q','Conductivity:Q','Organic_carbon:Q','Turbidity:Q']
).properties(title="pH vs. Turbidity").interactive()

On observing the plots of our dataset, we notice that there are pH values that do not make sense. From our research we know that the ideal pH range of water that is drinkable is between 6.5 - 8.5. However, this is not seen from the given dataset. For instance, the data states that a pH value of 4 is drinkable.There are also instances where the pH value is greater than 10 and is marked as drinkable.

corr_matrix = df.corr()
sns.heatmap(corr_matrix, cmap='coolwarm', annot=True, fmt='.2f', linewidths=0.5)

<AxesSubplot:>

1.2 Data Reduction

From the following box and violin plots, we can notice there are many outliers present in our dataset.

Outlier treatment - Since this is a synthetic data, they have no real significance and also because the features are uncorrelated, we will clip the outlier values to better fit the model. Reference - https://datascience.stackexchange.com/questions/65802/for-outliers-treatment-clipping-winsorizing-or-removing

Performing a 90% winsorization window

Q1 = df['ph'].quantile(0.05)
Q2 = df['ph'].quantile(0.95)
df['ph'] = df['ph'].clip(lower = Q1, upper = Q2)

Q1 = df['Hardness'].quantile(0.05)
Q2 = df['Hardness'].quantile(0.95)
df['Hardness'] = df['Hardness'].clip(lower = Q1, upper = Q2)

Q1 = df['Solids'].quantile(0.05)
Q2 = df['Solids'].quantile(0.95)
df['Solids'] = df['Solids'].clip(lower = Q1, upper = Q2)

Q1 = df['Chloramines'].quantile(0.05)
Q2 = df['Chloramines'].quantile(0.95)
df['Chloramines'] = df['Chloramines'].clip(lower = Q1, upper = Q2)

Q1 = df['Sulfate'].quantile(0.05)
Q2 = df['Sulfate'].quantile(0.95)
df['Sulfate'] = df['Sulfate'].clip(lower = Q1, upper = Q2)

Q1 = df['Conductivity'].quantile(0.05)
Q2 = df['Conductivity'].quantile(0.95)
df['Conductivity'] = df['Conductivity'].clip(lower = Q1, upper = Q2)

Q1 = df['Organic_carbon'].quantile(0.05)
Q2 = df['Organic_carbon'].quantile(0.95)
df['Organic_carbon'] = df['Organic_carbon'].clip(lower = Q1, upper = Q2)

Q1 = df['Trihalomethanes'].quantile(0.05)
Q2 = df['Trihalomethanes'].quantile(0.95)
df['Trihalomethanes'] = df['Trihalomethanes'].clip(lower = Q1, upper = Q2)

Q1 = df['Turbidity'].quantile(0.05)
Q2 = df['Turbidity'].quantile(0.95)
df['Turbidity'] = df['Turbidity'].clip(lower = Q1, upper = Q2)

#Saved outlier-free data
df.to_csv('Outlier_treated_data.csv')

from sklearn.preprocessing import StandardScaler

X1 = df.copy()
X1 = X1.drop('Potability',axis=1)

scaler = StandardScaler()
scaler.fit(X1)
Xa = scaler.transform(X1)
Xa[0:5]

array([[-3.31823713e-03,  2.95445357e-01, -1.39254296e-01,
         1.22913381e-01,  1.13490507e+00,  1.84182066e+00,
        -1.29603398e+00,  1.46127900e+00, -1.41117235e+00],
       [-1.87768215e+00, -1.88370849e+00, -4.09992788e-01,
        -3.48132686e-01, -2.14645385e-02,  1.86891055e+00,
         2.99743056e-01, -7.15347231e-01,  7.51401172e-01],
       [ 7.94661459e-01,  9.63263664e-01, -2.49713435e-01,
         1.52243217e+00, -2.14645385e-02, -9.23829495e-02,
         8.61105132e-01,  9.93726616e-04, -1.28064841e+00],
       [ 9.65209029e-01,  6.22796964e-01,  1.44628014e-02,
         6.60656263e-01,  7.55451534e-01, -8.27018833e-01,
         1.38232994e+00,  1.79872922e+00,  9.31598080e-01],
       [ 1.57287019e+00, -5.25750488e-01, -4.91551707e-01,
        -4.10927337e-01, -7.69839745e-01, -3.60477394e-01,
        -9.04257456e-01, -1.88117169e+00,  1.52807760e-01]])

from statsmodels.stats.outliers_influence import variance_inflation_factor
#dfq = pd.DataFrame(Xa,columns=df.columns)
dfq = pd.DataFrame(Xa,columns=X1.columns)
dfq= dfq.dropna()

# Select only the predictor variables you want to check for multicollinearity
predictors = dfq[['ph','Hardness','Solids','Chloramines','Sulfate','Conductivity','Organic_carbon','Trihalomethanes','Turbidity']]

# Create a numpy array from the predictor variables
X = np.array(predictors)

# Calculate the VIF values for each predictor variable
vif = [variance_inflation_factor(X, i) for i in range(X.shape[1])]

# Print the VIF values
print(vif)

[1.0199647537705459, 1.0217545715528653, 1.0325165306406208, 1.0075228023572673, 1.0293293618560084, 1.0026977082001023, 1.0036444941462002, 1.0023438017002047, 1.0039011393372756]

From this part we understand that there is no multicolinearity within the features.

dfq.head()

	ph	Hardness	Solids	Chloramines	Sulfate	Conductivity	Organic_carbon	Trihalomethanes	Turbidity
0	-0.003318	0.295445	-0.139254	0.122913	1.134905	1.841821	-1.296034	1.461279	-1.411172
1	-1.877682	-1.883708	-0.409993	-0.348133	-0.021465	1.868911	0.299743	-0.715347	0.751401
2	0.794661	0.963264	-0.249713	1.522432	-0.021465	-0.092383	0.861105	0.000994	-1.280648
3	0.965209	0.622797	0.014463	0.660656	0.755452	-0.827019	1.382330	1.798729	0.931598
4	1.572870	-0.525750	-0.491552	-0.410927	-0.769840	-0.360477	-0.904257	-1.881172	0.152808

X1 = X1.dropna()
X1

	ph	Hardness	Solids	Chloramines	Sulfate	Conductivity	Organic_carbon	Trihalomethanes	Turbidity
0	7.080795	204.890455	20791.318981	7.300212	368.516441	564.308654	10.379783	86.990970	2.963135
1	4.688853	141.763281	18630.057858	6.635246	333.073546	566.349320	15.180013	56.329076	4.500656
2	8.099124	224.236259	19909.541732	9.275884	333.073546	418.606213	16.868637	66.420093	3.055934
3	8.316766	214.373394	22018.417441	8.059332	356.886136	363.266516	18.436524	91.744595	4.628771
4	9.092223	181.101509	17978.986339	6.546600	310.135738	398.410813	11.558279	39.906235	4.075075
...	...	...	...	...	...	...	...	...	...
3271	4.688853	193.681735	38474.990249	7.166639	359.948574	526.424171	13.894419	66.687695	4.435821
3272	7.808856	193.553212	17329.802160	8.061362	333.073546	392.449580	19.637254	66.396293	2.798243
3273	9.419510	175.762646	33155.578218	7.350233	333.073546	432.044783	11.039070	69.845400	3.298875
3274	5.126763	230.603758	11983.869376	6.303357	333.073546	402.883113	11.168946	77.488213	4.708658
3275	7.874671	195.102299	17404.177061	7.509306	333.073546	327.459760	16.140368	78.698446	2.684279

3276 rows × 9 columns

2. Classification Modeling

2.1 Using Logistic Regression:

#Convert back for model building
def num_pot_label(p):
    if(p == 'No'):
        return 0
    else:
        return 1
    
df.Potability = df.Potability.apply(num_pot_label)
df.head()

	ph	Hardness	Solids	Chloramines	Sulfate	Conductivity	Organic_carbon	Trihalomethanes	Turbidity	Potability
0	7.080795	204.890455	20791.318981	7.300212	368.516441	564.308654	10.379783	86.990970	2.963135	0
1	4.688853	141.763281	18630.057858	6.635246	333.073546	566.349320	15.180013	56.329076	4.500656	0
2	8.099124	224.236259	19909.541732	9.275884	333.073546	418.606213	16.868637	66.420093	3.055934	0
3	8.316766	214.373394	22018.417441	8.059332	356.886136	363.266516	18.436524	91.744595	4.628771	0
4	9.092223	181.101509	17978.986339	6.546600	310.135738	398.410813	11.558279	39.906235	4.075075	0

#Blocks for data reduction in order to build models.
df1 = pd.read_csv("Outlier_treated_data.csv")
from sklearn.linear_model import LogisticRegression
from sklearn.model_selection import train_test_split
from sklearn.metrics import confusion_matrix
from sklearn.metrics import accuracy_score
X = df[["ph","Hardness","Solids","Chloramines","Sulfate","Conductivity","Organic_carbon","Trihalomethanes","Turbidity"]]
Y = df.Potability
X_train, X_test, Y_train, Y_test = train_test_split(X, Y, test_size = 0.25, random_state = 0)
log_reg = LogisticRegression()

log_reg.fit(X_train,Y_train)
y_pred = log_reg.predict(X_test)

print(y_pred[0:5],"\n")
print(Y_test.head())

[0 0 0 0 0] 

2017    1
2533    0
589     0
482     0
2620    0
Name: Potability, dtype: int64

conf_matrix = confusion_matrix(Y_test, y_pred)
  
print ("Confusion Matrix :\n", conf_matrix,"\n\n")
log_reg_accuracy = accuracy_score(Y_test, y_pred)

Confusion Matrix :
 [[502   0]
 [317   0]] 

sns.set(font_scale=1.4)
sns.heatmap(conf_matrix, annot=True, fmt='g', cmap='Blues')

plt.title('Confusion matrix of Logistic Regression')
plt.xlabel('Predicted')
plt.ylabel('True')
plt.show()
print (f"Model accuracy :{round(log_reg_accuracy,2)*100}%")

Model accuracy :61.0%

2.2 Using Decision Tree Classification

from sklearn.tree import DecisionTreeClassifier
from sklearn.model_selection import train_test_split
from sklearn.metrics import accuracy_score

dtree = DecisionTreeClassifier()
dtree.fit(X_train, Y_train)
dtree_y_pred = dtree.predict(X_test)

dtree_accuracy = accuracy_score(Y_test, dtree_y_pred)
print("Accuracy: {:.2f}%".format(dtree_accuracy * 100))

Accuracy: 55.31%

dtree_conf_matrix = confusion_matrix(Y_test, dtree_y_pred)
print(dtree_conf_matrix)

[[322 180]
 [186 131]]

sns.set(font_scale=1.4)
sns.heatmap(dtree_conf_matrix, annot=True, fmt='g', cmap='icefire')

plt.title('Confusion matrix of KNN classifier')
plt.xlabel('Predicted')
plt.ylabel('True')
plt.show()
print(f"Model Accuracy of KNN :{round(dtree_accuracy*100,2)}%")

Model Accuracy of KNN :55.31%

2.3 Using KNN Classifier

from sklearn.metrics import accuracy_score, confusion_matrix
from sklearn.neighbors import KNeighborsClassifier

knn = KNeighborsClassifier(n_neighbors=5)
knn.fit(X_train, Y_train)

y_pred = knn.predict(X_test)
knn_acc = accuracy_score(Y_test, y_pred)

# Create a confusion matrix
cm = confusion_matrix(Y_test, y_pred)
print(cm)

[[363 139]
 [230  87]]

sns.set(font_scale=1.4)
sns.heatmap(cm, annot=True, fmt='g', cmap='YlOrBr')

plt.title('Confusion matrix of KNN classifier')
plt.xlabel('Predicted')
plt.ylabel('True')
plt.show()
print(f"Model Accuracy of KNN :{round(knn_acc*100,2)}%")

Model Accuracy of KNN :54.95%

2.4 Using Random Forest Classifier

from sklearn.ensemble import RandomForestClassifier

# Create a Random Forest model with 100 trees
rf_model = RandomForestClassifier(n_estimators=100)

# Fit the model to the training data
rf_model.fit(X_train, Y_train)

# Make predictions on the test data
y_pred = rf_model.predict(X_test)

# Calculate the accuracy score
rf_acc = accuracy_score(Y_test, y_pred)

rf_y_pred = rf_model.predict(X_test)
rf_cm = confusion_matrix(Y_test, rf_y_pred)
print(rf_cm)

[[444  58]
 [218  99]]

sns.set(font_scale=1.4)
sns.heatmap(rf_cm, annot=True, fmt='g', cmap='coolwarm')

plt.title('Confusion matrix of Random Forest classifier')
plt.xlabel('Predicted')
plt.ylabel('True')
plt.show()
print(f"Model Accuracy of Random Forest is : {round(rf_acc*100,2)}%")

Model Accuracy of Random Forest is : 66.3%

import xgboost as xgb

xgb_model = xgb.XGBClassifier()

xgb_model.fit(X_train, Y_train)

y_pred = xgb_model.predict(X_test)

xg_acc = accuracy_score(Y_test, y_pred)

xg_y_pred = xgb_model.predict(X_test)
xg_cm = confusion_matrix(Y_test, y_pred)
print(xg_cm)

[[400 102]
 [192 125]]

sns.set(font_scale=1.4)
sns.heatmap(xg_cm, annot=True, fmt='g', cmap='rocket')

plt.title('Confusion matrix of XGBoost classifier')
plt.xlabel('Predicted')
plt.ylabel('True')
plt.show()
print(f"Model Accuracy of XGBoost : {round(xg_acc*100,2)}%")

Model Accuracy of XGBoost : 64.1%

Conclusion:

1. The dataset we are working with is Synthetic. The values in the corresponding columns have no real significance. Therefore, no research can be incorporated into our analysis.
2. The features in the dataset are uncorrelated to the dependent variables.
3. There is no multi-collinearity present among the features in the dataset.